Abstract: With the development of competition in business and also today’s rapid changes in competitive market, understanding these changes are the key factor of being more efficient in markets. CRM which is known as basic structure for describing customer’s needs for efficiently understand customer’s behavior and finely get the maximum of market share and profit. There are major differences between B2B and B2C businesses such as long term purchase cycle, purchase interests and the amount of the transactions. These differences needs more interactive strategies. The knowledge that gets from CRM is extremely related to market changes. In recent years data mining increasingly help organizations to get and understand customer’s behavior. But with the rapid changes in market these procedures must be change too. Change mining as the higher order of data mining tries to get knowledge by analysis patterns instead of data. In this paper we attempt to calculate customer’s value by using RFM model and K-MEANS clustering method and then analysis changes in deferent time periods. We tries to find out cluster transitions and most frequent customer value changing trend for proactive decision making. For this purpose we use customer purchase transactions in insurance industry which are gathered in 3 years.